Automated Brain Tumour Detection Using Deep Learning

Overview: This project develops an automated system for detecting brain tumors from MRI scans using a deep learning model. The goal is to assist radiologists in diagnosing brain tumors more accurately and efficiently, leveraging the capabilities of artificial intelligence.

Objective: The primary objective of this project is to improve the accuracy and speed of brain tumor detection processes in clinical settings. By automating the initial screening and diagnosis phases, the system aims to reduce human error and provide support in decision-making for medical professionals.

Methodology: The system utilizes a Convolutional Neural Network (CNN), specifically a pretrained ResNet-18 model, fine-tuned to identify various types of brain tumors in magnetic resonance imaging (MRI) data. The model was trained and validated on a comprehensive dataset of MRI images, ensuring robustness and reliability in diverse clinical environments.

Performance: The model achieved a remarkable accuracy of 97% in detecting brain tumors, demonstrating its effectiveness in distinguishing between tumorous and non-tumorous MRI scans. This high level of accuracy ensures that the system can be a reliable tool for preliminary assessments in medical diagnostics.

Implementation: To make the model accessible to medical practitioners, it is integrated into a web-based application. This application allows users to upload MRI images and receive immediate predictions regarding the presence of brain tumors. The user-friendly interface ensures that the system is accessible to medical staff with varying levels of technical expertise.

Future Work: Future enhancements will focus on expanding the dataset to include more varied types of brain tumors and implementing additional features such as tumor localization and classification by severity. Further research will also explore the integration of the model with hospital information systems to streamline the workflow in medical diagnostics.